Initial commit

2018-04-02 08:07:38 +02:00
commit 7330c1ed3e
2054 changed files with 405203 additions and 0 deletions
--- a/typo3conf/ext/dd_googlesitemap/Classes/Scheduler/AdditionalFieldsProvider.php
+++ b/typo3conf/ext/dd_googlesitemap/Classes/Scheduler/AdditionalFieldsProvider.php
@@ -0,0 +1,209 @@
+<?php
+/***************************************************************
+*  Copyright notice
+*
+*  (c) 2013-2014 Dmitry Dulepov <dmitry.dulepov@gmail.com>
+*  All rights reserved
+*
+*  This script is part of the TYPO3 project. The TYPO3 project is
+*  free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  The GNU General Public License can be found at
+*  http://www.gnu.org/copyleft/gpl.html.
+*
+*  This script is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  This copyright notice MUST APPEAR in all copies of the script!
+***************************************************************/
+
+namespace DmitryDulepov\DdGooglesitemap\Scheduler;
+
+use TYPO3\CMS\Core\Utility\GeneralUtility;
+
+/**
+ * This class provides information about additional fields for the scheduler
+ * task of this extension. Additional fields are:
+ * - the URL of the eID script (users can use different parameters for the script!)
+ * - index file path (users will submit that to Google)
+ * - maximum number of URLs in the sitemap (to prevent out of memory errors)
+ *
+ * WARNING! Due to incompatible TYPO3 6.2 changes this class now shows PHP errors
+ * in TYPO3 4.5 and TYPO3 6.2 in the PhpStorm. These errors however do not happen
+ * at runtime.
+ *
+ * @author Dmitry Dulepov <dmitry.dulepov@gmail.com>
+ */
+class AdditionalFieldsProvider implements \TYPO3\CMS\Scheduler\AdditionalFieldProviderInterface {
+
+	/**
+	 * Gets additional fields to render in the form to add/edit a task
+	 *
+	 * @param array $taskInfo Values of the fields from the add/edit task form
+	 * @param \TYPO3\CMS\Scheduler\Task\AbstractTask $task The task object being edited. Null when adding a task!
+	 * @param \TYPO3\CMS\Scheduler\Controller\SchedulerModuleController $schedulerModule Reference to the scheduler backend module
+	 * @return array A two dimensional array, array('Identifier' => array('fieldId' => array('code' => '', 'label' => '', 'cshKey' => '', 'cshLabel' => ''))
+	 */
+	public function getAdditionalFields(array &$taskInfo, $task, \TYPO3\CMS\Scheduler\Controller\SchedulerModuleController $schedulerModule) {
+		/** @var \DmitryDulepov\DdGooglesitemap\Scheduler\Task $task */
+		$additionalFields = array();
+
+		if (!$task) {
+			$url = GeneralUtility::locationHeaderUrl('/index.php?eID=dd_googlesitemap');
+			$task = GeneralUtility::makeInstance('DmitryDulepov\\DdGooglesitemap\\Scheduler\\Task');
+		}
+		else {
+			$url = $task->getEIdScriptUrl();
+		}
+		$indexFilePath = $task->getIndexFilePath();
+		$maxUrlsPerSitemap = $task->getMaxUrlsPerSitemap();
+
+		$additionalFields['eIdUrl'] = array(
+			'code'     => '<textarea style="width:350px;height:200px" name="tx_scheduler[eIdUrl]" wrap="off">' . htmlspecialchars($url) . '</textarea>',
+			'label'    => 'LLL:EXT:dd_googlesitemap/locallang.xml:scheduler.eIDFieldLabel',
+			'cshKey'   => '',
+			'cshLabel' => ''
+		);
+		$additionalFields['indexFilePath'] = array(
+			'code'     => '<input class="wide" type="text" name="tx_scheduler[indexFilePath]" value="' . htmlspecialchars($indexFilePath) . '" />',
+			'label'    => 'LLL:EXT:dd_googlesitemap/locallang.xml:scheduler.indexFieldLabel',
+			'cshKey'   => '',
+			'cshLabel' => ''
+		);
+		$additionalFields['maxUrlsPerSitemap'] = array(
+			'code'     => '<input type="text" name="tx_scheduler[maxUrlsPerSitemap]" value="' . $maxUrlsPerSitemap . '" />',
+			'label'    => 'LLL:EXT:dd_googlesitemap/locallang.xml:scheduler.maxUrlsPerSitemapLabel',
+			'cshKey'   => '',
+			'cshLabel' => ''
+		);
+
+		return $additionalFields;
+	}
+
+	/**
+	 * Validates the additional fields' values
+	 *
+	 * @param array $submittedData An array containing the data submitted by the add/edit task form
+	 * @param \TYPO3\CMS\Scheduler\Controller\SchedulerModuleController $schedulerModule Reference to the scheduler backend module
+	 * @return boolean TRUE if validation was ok (or selected class is not relevant), FALSE otherwise
+	 */
+	public function validateAdditionalFields(array &$submittedData, \TYPO3\CMS\Scheduler\Controller\SchedulerModuleController $schedulerModule) {
+		$errors = array();
+
+		$this->validateEIdUrl($submittedData, $errors);
+		$this->validateMaxUrlsPerSitemap($submittedData, $errors);
+		$this->validateIndexFilePath($submittedData, $errors);
+
+		foreach ($errors as $error) {
+			/** @noinspection PhpUndefinedMethodInspection */
+			$error = $GLOBALS['LANG']->sL('LLL:EXT:dd_googlesitemap/locallang.xml:' . $error);
+			$this->addErrorMessage($error);
+		}
+
+		return count($errors) == 0;
+	}
+
+	/**
+	 * Takes care of saving the additional fields' values in the task's object
+	 *
+	 * @param array $submittedData An array containing the data submitted by the add/edit task form
+	 * @param \TYPO3\CMS\Scheduler\Task\AbstractTask $task Reference to the scheduler backend module
+	 * @return void
+	 */
+	public function saveAdditionalFields(array $submittedData, \TYPO3\CMS\Scheduler\Task\AbstractTask $task) {
+		/** @var \DmitryDulepov\DdGooglesitemap\Scheduler\Task $task */
+		$task->setEIdScriptUrl($submittedData['eIdUrl']);
+		$task->setMaxUrlsPerSitemap($submittedData['maxUrlsPerSitemap']);
+		$task->setIndexFilePath($submittedData['indexFilePath']);
+	}
+
+	/**
+	 * Adds a error message as a flash message.
+	 *
+	 * @param string $message
+	 * @return void
+	 */
+	protected function addErrorMessage($message) {
+		$flashMessage = \TYPO3\CMS\Core\Utility\GeneralUtility::makeInstance('TYPO3\\CMS\\Core\\Messaging\\FlashMessage',
+			$message, '', \TYPO3\CMS\Core\Messaging\FlashMessage::ERROR
+		);
+		/** @var \TYPO3\CMS\Core\Messaging\FlashMessage $flashMessage */
+		$flashMessageService = \TYPO3\CMS\Core\Utility\GeneralUtility::makeInstance('TYPO3\\CMS\\Core\\Messaging\\FlashMessageService');
+		/** @var \TYPO3\CMS\Core\Messaging\FlashMessageService $flashMessageService */
+		$flashMessageService->getMessageQueueByIdentifier()->enqueue($flashMessage);
+	}
+
+	/**
+	 * Validates the number of urls per sitemap.
+	 *
+	 * @param array $submittedData
+	 * @param array $errors
+	 * @return void
+	 */
+	protected function validateMaxUrlsPerSitemap(array &$submittedData, array &$errors) {
+		$submittedData['maxUrlsPerSitemap'] = intval($submittedData['maxUrlsPerSitemap']);
+		if ($submittedData['maxUrlsPerSitemap'] <= 0) {
+			$errors[] = 'scheduler.error.badNumberOfUrls';
+		}
+	}
+
+	/**
+	 * Validates index file path.
+	 *
+	 * @param array $submittedData
+	 * @param array $errors
+	 * @return void
+	 */
+	protected function validateIndexFilePath(array &$submittedData, array &$errors) {
+		if (GeneralUtility::isAbsPath($submittedData['indexFilePath'])) {
+			$errors[] = 'scheduler.error.badIndexFilePath';
+		}
+		else {
+			$testPath = GeneralUtility::getFileAbsFileName($submittedData['indexFilePath'], TRUE);
+			if (!file_exists($testPath)) {
+				if (!@touch($testPath)) {
+					$errors[] = 'scheduler.error.badIndexFilePath';
+				}
+				else {
+					unlink($testPath);
+				}
+			}
+		}
+	}
+
+	/**
+	 * Valies the URL of the eID script.
+	 *
+	 * @param array $submittedData
+	 * @param array $errors
+	 */
+	protected function validateEIdUrl(array &$submittedData, array &$errors) {
+		foreach (GeneralUtility::trimExplode(chr(10), $submittedData['eIdUrl']) as $url) {
+			if (FALSE !== ($urlParts = parse_url($url))) {
+				if (!$urlParts['host']) {
+					$errors[] = 'scheduler.error.missingHost';
+				}
+				else {
+					/** @noinspection PhpUndefinedMethodInspection */
+					list($count) = $GLOBALS['TYPO3_DB']->exec_SELECTgetRows('COUNT(*) AS counter', 'sys_domain',
+							'domainName=' . $GLOBALS['TYPO3_DB']->fullQuoteStr($urlParts['host'], 'sys_domain')
+					);
+					if ($count['counter'] == 0) {
+						$errors[] = 'scheduler.error.missingHost';
+					}
+				}
+				if (!preg_match('/(?:^|&)eID=dd_googlesitemap/', $urlParts['query'])) {
+					$errors[] = 'scheduler.error.badPath';
+				}
+				if (preg_match('/(?:^|&)(?:offset|limit)=/', $urlParts['query'])) {
+					$errors[] = 'scheduler.error.badParameters';
+				}
+			}
+		}
+	}
+}
--- a/typo3conf/ext/dd_googlesitemap/Classes/Scheduler/Task.php
+++ b/typo3conf/ext/dd_googlesitemap/Classes/Scheduler/Task.php
@@ -0,0 +1,300 @@
+<?php
+/***************************************************************
+*  Copyright notice
+*
+*  (c) 2013 Dmitry Dulepov <dmitry.dulepov@gmail.com>
+*  All rights reserved
+*
+*  This script is part of the TYPO3 project. The TYPO3 project is
+*  free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  The GNU General Public License can be found at
+*  http://www.gnu.org/copyleft/gpl.html.
+*
+*  This script is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  This copyright notice MUST APPEAR in all copies of the script!
+***************************************************************/
+
+namespace DmitryDulepov\DdGooglesitemap\Scheduler;
+
+use TYPO3\CMS\Core\Utility\GeneralUtility;
+
+/**
+ * This class provides a scheduler task to create sitemap index as required
+ * by the Google sitemap protocol.
+ *
+ * @author Dmitry Dulepov <dmitry.dulepov@gmail.com>
+ * @see http://support.google.com/webmasters/bin/answer.py?hl=en&answer=71453
+ */
+class Task extends \TYPO3\CMS\Scheduler\Task\AbstractTask {
+
+	const DEFAULT_FILE_PATH = 'typo3temp/dd_googlesitemap';
+
+	/** @var string */
+	private $baseUrl;
+
+	/** @var string */
+	protected $eIdScriptUrl;
+
+	/** @var string */
+	protected $indexFilePath;
+
+	/** @var int */
+	protected $maxUrlsPerSitemap = 50000;
+
+	/** @var string */
+	private $sitemapFileFormat;
+
+	/** @var int */
+	private $offset;
+
+	/**
+	 * Creates the instance of the class. This call initializes the index file
+	 * path to the random value. After the task is configured, the user may
+	 * change the file and the file name will be serialized with the task and
+	 * used later.
+	 *
+	 * @see __sleep
+	 */
+	public function __construct() {
+		parent::__construct();
+		$this->indexFilePath = self::DEFAULT_FILE_PATH . '/' . GeneralUtility::getRandomHexString(24) . '.xml';
+	}
+
+	/**
+	 * Reconstructs some variables after the object is unserialized.
+	 *
+	 * @return void
+	 */
+	public function __wakeup() {
+		$this->buildSitemapFileFormat();
+		$this->buildBaseUrl();
+	}
+
+	/**
+	 * This is the main method that is called when a task is executed
+	 * It MUST be implemented by all classes inheriting from this one
+	 * Note that there is no error handling, errors and failures are expected
+	 * to be handled and logged by the client implementations.
+	 * Should return true on successful execution, false on error.
+	 *
+	 * @return boolean    Returns true on successful execution, false on error
+	 */
+	public function execute() {
+		$indexFilePathTemp = PATH_site . $this->indexFilePath . '.tmp';
+		$indexFile = fopen($indexFilePathTemp, 'wt');
+		fwrite($indexFile, '<?xml version="1.0" encoding="UTF-8"?>' . chr(10));
+		fwrite($indexFile, '<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">' . chr(10));
+
+		$eIDscripts = GeneralUtility::trimExplode(chr(10), $this->eIdScriptUrl);
+		$eIdIndex = 1;
+		foreach ($eIDscripts as $eIdScriptUrl) {
+			$this->offset = 0;
+			$currentFileNumber = 1;
+			$lastFileHash = '';
+			do {
+				$sitemapFileName = sprintf($this->sitemapFileFormat, $eIdIndex, $currentFileNumber++);
+				$this->buildSitemap($eIdScriptUrl, $sitemapFileName);
+
+				$isSitemapEmpty = $this->isSitemapEmpty($sitemapFileName);
+				$currentFileHash = $isSitemapEmpty ? -1 : md5_file(PATH_site . $sitemapFileName);
+				$stopLoop = $isSitemapEmpty || ($currentFileHash == $lastFileHash);
+
+				if ($stopLoop) {
+					@unlink(PATH_site . $sitemapFileName);
+				}
+				else {
+					fwrite($indexFile, '<sitemap><loc>' . htmlspecialchars($this->makeSitemapUrl($sitemapFileName)) . '</loc></sitemap>' . chr(10));
+					$lastFileHash = $currentFileHash;
+				}
+			} while (!$stopLoop);
+			$eIdIndex++;
+		}
+
+		fwrite($indexFile, '</sitemapindex>' . chr(10));
+		fclose($indexFile);
+
+		@unlink(PATH_site . $this->indexFilePath);
+		rename($indexFilePathTemp, PATH_site . $this->indexFilePath);
+
+		return true;
+	}
+
+	/**
+	 * This method is designed to return some additional information about the task,
+	 * that may help to set it apart from other tasks from the same class
+	 * This additional information is used - for example - in the Scheduler's BE module
+	 * This method should be implemented in most task classes
+	 *
+	 * @return	string	Information to display
+	 */
+	public function getAdditionalInformation() {
+		/** @noinspection PhpUndefinedMethodInspection */
+		$format = $GLOBALS['LANG']->sL('LLL:EXT:dd_googlesitemap/locallang.xml:scheduler.extra_info');
+		return sprintf($format, $this->getIndexFileUrl());
+	}
+
+	/**
+	 * Sets the url of the eID script. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @return string
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function getEIdScriptUrl() {
+		return $this->eIdScriptUrl;
+	}
+
+	/**
+	 * Returns the index file path. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @return string
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function getIndexFilePath() {
+		return $this->indexFilePath;
+	}
+
+	/**
+	 * Obtains the number of urls per sitemap. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @return int
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function getMaxUrlsPerSitemap() {
+		return $this->maxUrlsPerSitemap;
+	}
+
+	/**
+	 * Sets the URl of the eID script. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @param $url
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function setEIdScriptUrl($url) {
+		$this->eIdScriptUrl = $url;
+	}
+
+	/**
+	 * Sets the URL of the eID script. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @param string $path
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function setIndexFilePath($path) {
+		$this->indexFilePath = $path;
+	}
+
+	/**
+	 * Sets the number of URLs per sitemap. This is called from the task
+	 * configuration inside scheduler.
+	 *
+	 * @param int $maxUrlsPerSitemap
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	public function setMaxUrlsPerSitemap($maxUrlsPerSitemap) {
+		$this->maxUrlsPerSitemap = $maxUrlsPerSitemap;
+	}
+
+	/**
+	 * Creates a base url for sitemaps.
+	 *
+	 * @return void
+	 */
+	protected function buildBaseUrl() {
+		$urlParts = parse_url($this->eIdScriptUrl);
+		$this->baseUrl = $urlParts['scheme'] . '://';
+		if ($urlParts['user']) {
+			$this->baseUrl .= $urlParts['user'];
+			if ($urlParts['pass']) {
+				$this->baseUrl .= ':' . $urlParts['pass'];
+			}
+			$this->baseUrl .= '@';
+		}
+		$this->baseUrl .= $urlParts['host'];
+		if ($urlParts['port']) {
+			$this->baseUrl .= ':' . $urlParts['port'];
+		}
+		$this->baseUrl .= '/';
+	}
+
+	/**
+	 * Builds the sitemap.
+	 *
+	 * @param string $eIdScriptUrl
+	 * @param string $sitemapFileName
+	 * @see tx_ddgooglesitemap_additionalfieldsprovider
+	 */
+	protected function buildSitemap($eIdScriptUrl, $sitemapFileName) {
+		$url = $eIdScriptUrl . sprintf('&offset=%d&limit=%d', $this->offset, $this->maxUrlsPerSitemap);
+
+		$content = GeneralUtility::getURL($url);
+		if ($content) {
+			file_put_contents(PATH_site . $sitemapFileName, $content);
+			$this->offset += $this->maxUrlsPerSitemap;
+		}
+	}
+
+	/**
+	 * Creates the format string for the sitemap files.
+	 *
+	 * @return void
+	 */
+	protected function buildSitemapFileFormat() {
+		$fileParts = pathinfo($this->indexFilePath);
+		$this->sitemapFileFormat = $fileParts['dirname'] . '/' . $fileParts['filename'] . '_sitemap_%05d_%05d.xml';
+	}
+
+	/**
+	 * Returns the index file url.
+	 *
+	 * @return string
+	 */
+	protected function getIndexFileUrl() {
+		return $this->baseUrl . $this->indexFilePath;
+	}
+
+	/**
+	 * Checks if the current sitemap has no entries. The function reads a chunk
+	 * of the file, which is large enough to have a '<url>' token in it and
+	 * examines the chunk. If the token is not found, than the sitemap is either
+	 * empty or corrupt.
+	 *
+	 * @param string $sitemapFileName
+	 * @return bool
+	 */
+	protected function isSitemapEmpty($sitemapFileName) {
+		$result = TRUE;
+
+		$fileDescriptor = @fopen(PATH_site . $sitemapFileName, 'rt');
+		if ($fileDescriptor) {
+			$chunkSizeToCheck = 10240;
+			$testString = fread($fileDescriptor, $chunkSizeToCheck);
+			fclose($fileDescriptor);
+			$result = (strpos($testString, '<url>') === FALSE);
+		}
+
+		return $result;
+	}
+
+	/**
+	 * Creates a url to the sitemap.
+	 *
+	 * @param string $siteMapPath
+	 * @return string
+	 */
+	protected function makeSitemapUrl($siteMapPath) {
+		return $this->baseUrl . $siteMapPath;
+	}
+}